:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
// services/mcp/tests/unit/schema.ajv.test.ts
// Guard test: compiles MCP tool input schemas with AJV in strict mode.
// Uses the generated schema at services/mcp/schema/tool-inputs.json to catch
// invalid or Ajv-strict-incompatible changes early in CI.
import Ajv from 'ajv'
import addFormats from 'ajv-formats'
import { readFile } from 'node:fs/promises'
import path from 'node:path'
import { fileURLToPath } from 'node:url'
import { expect, test } from 'vitest'
test('AJV strict compiles all MCP tool input schemas', async () => {
// Resolve ../../schema/tool-inputs.json relative to this test file
const __filename = fileURLToPath(import.meta.url)
const __dirname = path.dirname(__filename)
const schemaPath = path.resolve(__dirname, '../../schema/tool-inputs.json')
let doc: unknown
try {
const src = await readFile(schemaPath, 'utf8')
doc = JSON.parse(src)
} catch {
// Guidance if the generated schema file is missing
throw new Error(
`Could not read schema file at ${schemaPath}.
If this package generates schemas, run: pnpm --dir services/mcp schema:build:json`
)
}
// Support multiple shapes:
// - { definitions: { [name]: schema } } (local generated schema)
// - { tools: { [toolName]: schema } }
// - { tools: Array<{ name, inputSchema }> }
// - Plain array of { name, inputSchema }
// - { [toolName]: schema }
const extract = (value: any): Array<{ name: string; schema: any }> => {
if (!value) {
return []
}
if (Array.isArray(value)) {
return value
.map((t) => ('inputSchema' in t ? { name: t.name ?? 'unknown', schema: t.inputSchema } : null))
.filter(Boolean) as Array<{ name: string; schema: any }>
}
if (value.definitions) {
return extract(value.definitions)
}
if (value.tools) {
return extract(value.tools)
}
if (typeof value === 'object') {
return Object.entries(value).map(([name, schema]) => ({ name, schema }))
}
return []
}
const items = extract(doc)
expect(items.length).toBeGreaterThan(0)
// Ajv strict mode surfaces schema issues (e.g. duplicate enum values, invalid unions)
const ajv = new Ajv({
strict: true,
allErrors: true,
allowUnionTypes: true,
})
addFormats(ajv)
const failures: Array<{ name: string; error: unknown }> = []
for (const { name, schema } of items) {
try {
ajv.compile(schema)
} catch (err) {
failures.push({ name, error: err })
}
}
// If this assertion fails, at least one tool input schema is invalid under Ajv strict mode.
expect(failures, JSON.stringify(failures, null, 2)).toHaveLength(0)
})